In [2]:
import pandas as pd
import numpy as np
import matplotlib as mat
import seaborn as sb
import sklearn as sk
%matplotlib inline
Reading file¶
In [3]:
df=pd.read_csv("ShoeStore_Feedback.csv")
df.head()
Out[3]:
In [4]:
df.info()
Checking how many rows has null values in the primary Question Column¶
In [5]:
df[df['Overall Experience Score'].isnull()].info()
Dropped rows with null values in primary Question column.¶
In [6]:
df2=df.dropna(subset=['Overall Experience Score'],axis=0)
In [7]:
df2.info()
I think these two rows are to test the system if it is working¶
In [9]:
df2[df2['Storeid'].isnull()]
Out[9]:
Count of group merchandizer¶
In [10]:
df2['Group Merchandizer'].value_counts()
Out[10]:
Count of city managers¶
In [11]:
df2['Group CityManager'].value_counts()
Out[11]:
Count of Regional heads¶
In [12]:
df2['Group Regional_Head'].value_counts()
Out[12]:
Business head¶
In [13]:
df2['Group Business_Head'].value_counts()
Out[13]:
Count of group manager¶
In [14]:
df2['Group Management'].value_counts()
Out[14]:
Dropping irrelevant data¶
In [15]:
df2=df2.drop('Group Business_Head',axis=1)
In [16]:
df2=df2.drop('Group Management',axis=1)
In [17]:
df2=df2.drop('Group Regional_Head',axis=1)
In [18]:
df2=df2.drop('Group CityManager',axis=1)
In [19]:
df2=df2.drop('Group Merchandizer',axis=1)
Language count¶
In [20]:
df2['Language'].value_counts()
Out[20]:
Count of product category¶
In [21]:
df2['Product Category'].value_counts()
Out[21]:
In [90]:
df2['Product Category'].nunique()
Out[90]:
Count of different store attributes¶
In [22]:
df2['Store Ambience'].value_counts()
Out[22]:
In [23]:
df2['Store Collection'].value_counts()
Out[23]:
In [24]:
df2['Staff Product Knowledge'].value_counts()
Out[24]:
In [25]:
df2['Staff Helpfulness'].value_counts()
Out[25]:
Mapping the store attributes to integers¶
In [26]:
df2['Store Ambience']=df2['Store Ambience'].map({'Very Good(4)':4,'Excellent(5)':5,'Good(3)':3,'Fair(2)':2,'Poor(1)':1 })
In [27]:
df2['Store Collection']=df2['Store Collection'].map({'Very Good(4)':4,'Excellent(5)':5,'Good(3)':3,'Fair(2)':2,'Poor(1)':1 })
In [28]:
df2['Staff Product Knowledge']=df2['Staff Product Knowledge'].map({'Very Good(4)':4,'Excellent(5)':5,'Good(3)':3,'Fair(2)':2,'Poor(1)':1 })
In [29]:
df2['Staff Helpfulness']=df2['Staff Helpfulness'].map({'Very Good(4)':4,'Excellent(5)':5,'Good(3)':3,'Fair(2)':2,'Poor(1)':1 })
Normalizing Value¶
In [30]:
df2['Log Value']=np.log(df2['Value'])
Plotting graphs¶
In [85]:
df2.hist(bins=100,figsize=(20,20))
Out[85]:
In [32]:
df2[['Store Ambience','Store Collection','Staff Product Knowledge','Staff Helpfulness']].plot.hist(stacked=True, bins=100,figsize=(20,20))
Out[32]:
In [88]:
mat.pyplot.scatter(df2['Overall Experience Score'],df2['Value'])
Out[88]:
In [89]:
mat.pyplot.scatter(df2['Overall Experience Score'],df2['Log Value'])
Out[89]:
In [92]:
sb.heatmap(df2[['Store Ambience','Store Collection','Staff Product Knowledge','Staff Helpfulness']].corr())
Out[92]:
In [33]:
import nltk
from wordcloud import WordCloud
from nltk.corpus import stopwords
st=df2['Comment'].to_string()
tokens=st.split()
stop = set(stopwords.words('english'))
resultwords = [word for word in tokens if word.lower() not in stop]
result = ' '.join(resultwords)
result=''.join([i for i in result if not i.isdigit()])
l=result.split()
try:
[i for i in l if str(i) != 'nan']
except:
{}
try:
for word in list(l):
l.remove(u'NaN')
except:
{}
s=' '.join(l)
whitelist = set('abcdefghijklmnopqrstuvwxy ABCDEFGHIJKLMNOPQRSTUVWXYZ')
answer = ''.join(filter(whitelist.__contains__,s))
wordcloud = WordCloud(background_color="white",
width=1200, height=1000).generate(answer)
mat.pyplot.figure(figsize=(30,100))
mat.pyplot.imshow(wordcloud, interpolation="bilinear")
mat.pyplot.axis("off")
mat.pyplot.margins(x=0, y=0)
mat.pyplot.show()
Count of different store locations¶
In [34]:
df2['Store Location(Touchpoint name)'].nunique()
Out[34]:
Counts of store at each location¶
In [35]:
df2['Store Location(Touchpoint name)'].value_counts()
Out[35]:
Analysis for locations which have top 5 feedbacks in terms of customers who have reviewed¶
Analysis for Store Location 'CLB-COLABA'¶
In [36]:
df3=df2.loc[df2['Store Location(Touchpoint name)'] == 'CLB-COLABA']
df3.head()
Out[36]:
In [37]:
df3['Log Value']=np.log(df3['Value'])
In [38]:
df3.hist(bins=100,figsize=(20,20))
Out[38]:
In [39]:
df3[['Store Ambience','Store Collection','Staff Product Knowledge','Staff Helpfulness']].plot.hist(stacked=True, bins=100,figsize=(15,20))
Out[39]:
In [40]:
st=df3['Comment'].to_string()
tokens=st.split()
stop = set(stopwords.words('english'))
resultwords = [word for word in tokens if word.lower() not in stop]
result = ' '.join(resultwords)
result=''.join([i for i in result if not i.isdigit()])
l=result.split()
try:
[i for i in l if str(i) != 'nan']
except:
{}
try:
for word in list(l):
l.remove(u'NaN')
except:
{}
s=' '.join(l)
whitelist = set('abcdefghijklmnopqrstuvwxy ABCDEFGHIJKLMNOPQRSTUVWXYZ')
answer = ''.join(filter(whitelist.__contains__,s))
wordcloud = WordCloud(background_color="white",
width=1200, height=1000).generate(answer)
mat.pyplot.figure(figsize=(30,100))
mat.pyplot.imshow(wordcloud, interpolation="bilinear")
mat.pyplot.axis("off")
mat.pyplot.margins(x=0, y=0)
mat.pyplot.show()
Analysis for Store Location 'HYD-HYDERABAD'¶
In [41]:
df4=df2.loc[df2['Store Location(Touchpoint name)'] == 'HYD-HYDERABAD']
df4.head()
Out[41]:
In [43]:
df4['Log Value']=np.log(df4['Value'])
In [44]:
df4.hist(bins=100,figsize=(20,20))
Out[44]:
In [45]:
df4[['Store Ambience','Store Collection','Staff Product Knowledge','Staff Helpfulness']].plot.hist(stacked=True, bins=100,figsize=(15,20))
Out[45]:
In [46]:
st=df4['Comment'].to_string()
tokens=st.split()
stop = set(stopwords.words('english'))
resultwords = [word for word in tokens if word.lower() not in stop]
result = ' '.join(resultwords)
result=''.join([i for i in result if not i.isdigit()])
l=result.split()
try:
[i for i in l if str(i) != 'nan']
except:
{}
try:
for word in list(l):
l.remove(u'NaN')
except:
{}
s=' '.join(l)
whitelist = set('abcdefghijklmnopqrstuvwxy ABCDEFGHIJKLMNOPQRSTUVWXYZ')
answer = ''.join(filter(whitelist.__contains__,s))
wordcloud = WordCloud(background_color="white",
width=1200, height=1000).generate(answer)
mat.pyplot.figure(figsize=(30,100))
mat.pyplot.imshow(wordcloud, interpolation="bilinear")
mat.pyplot.axis("off")
mat.pyplot.margins(x=0, y=0)
mat.pyplot.show()
Analysis for Store Location 'AAS-AHMEDABAD'¶
In [47]:
df5=df2.loc[df2['Store Location(Touchpoint name)'] == 'AAS-AHMEDABAD']
df5.head()
Out[47]:
In [48]:
df5['Log Value']=np.log(df5['Value'])
In [49]:
df5.hist(bins=100,figsize=(20,20))
Out[49]:
In [50]:
df5[['Store Ambience','Store Collection','Staff Product Knowledge','Staff Helpfulness']].plot.hist(stacked=True, bins=100,figsize=(15,20))
Out[50]:
In [51]:
st=df5['Comment'].to_string()
tokens=st.split()
stop = set(stopwords.words('english'))
resultwords = [word for word in tokens if word.lower() not in stop]
result = ' '.join(resultwords)
result=''.join([i for i in result if not i.isdigit()])
l=result.split()
try:
[i for i in l if str(i) != 'nan']
except:
{}
try:
for word in list(l):
l.remove(u'NaN')
except:
{}
s=' '.join(l)
whitelist = set('abcdefghijklmnopqrstuvwxy ABCDEFGHIJKLMNOPQRSTUVWXYZ')
answer = ''.join(filter(whitelist.__contains__,s))
wordcloud = WordCloud(background_color="white",
width=1200, height=1000).generate(answer)
mat.pyplot.figure(figsize=(30,100))
mat.pyplot.imshow(wordcloud, interpolation="bilinear")
mat.pyplot.axis("off")
mat.pyplot.margins(x=0, y=0)
mat.pyplot.show()
Analysis for Store Location 'CTR-COIMBATORE'¶
In [52]:
df6=df2.loc[df2['Store Location(Touchpoint name)'] == 'CTR-COIMBATORE']
df6.head()
Out[52]:
In [53]:
df6['Log Value']=np.log(df6['Value'])
In [54]:
df6.hist(bins=100,figsize=(20,20))
Out[54]:
In [55]:
df6[['Store Ambience','Store Collection','Staff Product Knowledge','Staff Helpfulness']].plot.hist(stacked=True, bins=100,figsize=(15,20))
Out[55]:
In [83]:
st=df6['Comment'].to_string()
tokens=st.split()
stop = set(stopwords.words('english'))
resultwords = [word for word in tokens if word.lower() not in stop]
result = ' '.join(resultwords)
result=''.join([i for i in result if not i.isdigit()])
l=result.split()
try:
[i for i in l if str(i) != 'nan']
except:
{}
try:
for word in list(l):
l.remove(u'NaN')
except:
{}
s=' '.join(l)
whitelist = set('abcdefghijklmnopqrstuvwxy ABCDEFGHIJKLMNOPQRSTUVWXYZ')
answer = ''.join(filter(whitelist.__contains__,s))
wordcloud = WordCloud(background_color="white",
width=1200, height=1000).generate(answer)
mat.pyplot.figure(figsize=(30,100))
mat.pyplot.imshow(wordcloud, interpolation="bilinear")
mat.pyplot.axis("off")
mat.pyplot.margins(x=0, y=0)
mat.pyplot.show()
Analysis for Store Location 'AND-ANDHERI'¶
In [56]:
df7=df2.loc[df2['Store Location(Touchpoint name)'] == 'AND-ANDHERI']
df7.head()
Out[56]:
In [58]:
df7['Log Value']=np.log(df7['Value'])
In [59]:
df7.hist(bins=100,figsize=(20,20))
Out[59]:
In [60]:
df7[['Store Ambience','Store Collection','Staff Product Knowledge','Staff Helpfulness']].plot.hist(stacked=True, bins=100,figsize=(15,20))
Out[60]:
In [84]:
st=df6['Comment'].to_string()
tokens=st.split()
stop = set(stopwords.words('english'))
resultwords = [word for word in tokens if word.lower() not in stop]
result = ' '.join(resultwords)
result=''.join([i for i in result if not i.isdigit()])
l=result.split()
try:
[i for i in l if str(i) != 'nan']
except:
{}
try:
for word in list(l):
l.remove(u'NaN')
except:
{}
s=' '.join(l)
whitelist = set('abcdefghijklmnopqrstuvwxy ABCDEFGHIJKLMNOPQRSTUVWXYZ')
answer = ''.join(filter(whitelist.__contains__,s))
wordcloud = WordCloud(background_color="white",
width=1200, height=1000).generate(answer)
mat.pyplot.figure(figsize=(30,100))
mat.pyplot.imshow(wordcloud, interpolation="bilinear")
mat.pyplot.axis("off")
mat.pyplot.margins(x=0, y=0)
mat.pyplot.show()
Analysis for locations which have bottom 5 feedbacks in terms of customers who have reviewed¶
Analysis of store location "BST-Bangaluru"¶
In [61]:
df8=df2.loc[df2['Store Location(Touchpoint name)'] == 'BST-BANGALURU']
df8.head()
Out[61]:
In [62]:
df8.hist(bins=100,figsize=(20,20))
Out[62]:
In [63]:
df8[['Store Ambience','Store Collection','Staff Product Knowledge','Staff Helpfulness']].plot.hist(stacked=True, bins=100,figsize=(15,20))
Out[63]:
In [64]:
df8['Comment']
Out[64]:
In [65]:
pd.set_option('display.max_colwidth', -1)
Analysis of store location "ALR-ALWAR"¶
In [66]:
df9=df2.loc[df2['Store Location(Touchpoint name)'] == 'ALR-ALWAR']
df9.head()
Out[66]:
In [67]:
df9=df2.loc[df2['Store Location(Touchpoint name)'] == 'ALR-ALWAR']
df9.head()
Out[67]:
In [68]:
df9.hist(bins=100,figsize=(20,20))
Out[68]:
In [69]:
df9[['Store Ambience','Store Collection','Staff Product Knowledge','Staff Helpfulness']].plot.hist(stacked=True, bins=100,figsize=(15,20))
Out[69]:
In [70]:
df9['Comment']
Out[70]:
Analysis of store location "MAS-MUMBAI AIRPORT(M)"¶
In [71]:
df10=df2.loc[df2['Store Location(Touchpoint name)'] == 'MAS- MUMBAI AIR PORT( M )']
df10.head()
Out[71]:
In [72]:
df10.hist(bins=100,figsize=(20,20))
Out[72]:
In [73]:
df10[['Store Ambience','Store Collection','Staff Product Knowledge','Staff Helpfulness']].plot.hist(stacked=True, bins=100,figsize=(15,20))
Out[73]:
In [74]:
df10['Comment']
Out[74]:
Analysis of store location "JHR-JAMSHEDPUR"¶
In [75]:
df11=df2.loc[df2['Store Location(Touchpoint name)'] == 'JHR- JAMSHEDPUR']
df11.head()
Out[75]:
In [76]:
df11.hist(bins=100,figsize=(20,20))
Out[76]:
In [77]:
df11[['Store Ambience','Store Collection','Staff Product Knowledge','Staff Helpfulness']].plot.hist(stacked=True, bins=100,figsize=(15,20))
Out[77]:
In [78]:
df11['Comment']
Out[78]:
Analysis of store location "BHR-BANGALURU"¶
In [79]:
df12=df2.loc[df2['Store Location(Touchpoint name)'] == 'BHR-BANGALURU']
df12.head()
Out[79]:
In [80]:
df12.hist(bins=100,figsize=(20,20))
Out[80]:
In [81]:
df12[['Store Ambience','Store Collection','Staff Product Knowledge','Staff Helpfulness']].plot.hist(stacked=True, bins=100,figsize=(15,20))
Out[81]:
In [82]:
df12['Comment']
Out[82]: